Aeromagnetic Data Interpretation for

the McDougal Lakes Area, Duluth Complex

Lake County, Minnesota

June, 1989

For The Minnesota Dept. of Natural Resources

By Robert J. Ferderer

Eagan, Minnesota

Table of Contents

Page

Objective . . ........................................... 1

Data ..................... . .......................... . . 1

Approach ................... . ............ . ............. 5

Discussion ............... . ..... . ..................... 10

Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Ref ereilces Cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 7

List of Figures

Page

Fig. 1 Location map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Fig. 2 Remanent magnetization example . ............... 4

Fig. 3 Ideal Werner deconvolution anomaly sources .... 6 . . . .

Fig. 4 Real anomaly sources that may be represented by Werner deconvolution results ... 7

Obj ective

In this study, aeromagnetic and to a lesser degree, gravity

data are analyzed for a four-township sized area in the

block T59-61N, R9-11W, Lake County, Minnesota (Fig. 1).

Techniques utilized include: Werner deconvolution-based

inverse modeling, Talwani-based forward modeling and

frequency-domain filtering. The objective of this study is

to define geologic structures and lithologic units within

the Duluth Complex, based on information provided by these

techniques.

Data

Digital aeromagnetic and gravity data were taken from

magnetic tapes owned by the Minnesota Geological Survey.

Aeromagnetic Data

The high-resolution aeromagnetic survey used in this study

was flown in 1979-80, and consists of flight lines oriented

south-north, spaced 400 m apart, and sampled at a 50 m

inte.rval. Tie lines are oriented west-east, spaced 2000 m

apart and sampled every 50 m. Both flight and tie lines

were flown at a mean terrain clearance of 150 m.

Aeromagnetic anomalies in the study area have amplitudes

ranging from < -2000 to > 1000 gammas. The anomaly

expression is complicated by strong remanent magnetization,

having an average declination and inclination of (290°, 400)

(Halls and Pesonen, 1982). Induced magnetization is assumed

to be aligned along the main geomagnetic field which has an

approximate declination and inclination of (3°, 75°). The

Koenigsberger (Q) ratio, or ratio of remanent to induced

magnetization generally ranges between 1 and 10 throughout

the study area (Holst and others, 1986).

The strong remanent magnetization described above results in

magnetic anomalies that are significantly asymmetrical. For

a west-east (south-north) profile, and a vertical prism

l

, ;:STERN u:.llto STAT ES l l~'oooo

.,

. ,

Figure 1.

T\YO lL-\H,BDl~S

Location map.

,__;_ I

/

/ . . / · ·

. , ·· . / ·

') ,_

-! . I

/ ... . / .. -r­

/ .

anomaly source, an anomaly minimum oc;curs near the western

(northern) contact and a maximum occurs near the eastern

(southern) contact (Fig. 2).

Gravity Data

Gravity coverage over the study area is poor, as

measurements have been taken at less than one hundred

stations. Bouguer gravity anomaly data indicate a strong

near-linear gradient, with anomaly values increasing from

northwest to southeast. This regional anomaly is

interpreted to be related to eastward thickening of the

Duluth Complex (Ferderer, 1982). Superimposed on the

regional anomaly are shorter-wavelength anomalies that are

more indicative of the near-surface geology in the study

area. The strongest of these anomalies has an amplitude of

about 20 milligals.

Rock Property Data

Rock property data pertinent to this study are summarized in

Table 1. NRM is·: the intensity of natural remanent

magnetization.

Ferrogabbro

B. Eagle Ring

Avg. Density (gm/cc)

3.18*

Troctolite 2.97

B. Eagle Core Gabbro 2.98

Anorthositic Rocks

Basal Troctolites

2.80

2.91

Avg. Susc. (cgs)

Avg. NRM (cgs)

.0050* .0058*

. 0007 . 0012

.0003 .0038

. 0010* . 0006*

. 0013 . 0033

Avg. Q

(cgs)

1. 9*

2. 8

21. 0

1. O*

4.2

Avg. NRM decl.

304*

303

291

296*

284

Avg. NRM incl.

32*

55

30

37*

44

Table 1. Rock property information for Duluth Complex rocks. Taken from Holst and others (1986), and .v.w. Chandler (personal communication). Asterisks denote values based on less than 5 measurements.

3

~

0 c c 0 0

0

.. c g r

c "'

0.. .. "'

We.st Eais-c

11Qgn~t ICS.

1500.00 /\ I

-1000 oa~--------------~~~---~~~~~-'

-o 00.....-----------~-----~---------~

5 . 00 +---.....---.....---.....---~--~--~--..---~--..---0 . 00 5 . 00 10 . 00

Distance (Kn1

. c c a

"'

~ c

"' . :i::

c

"' .J:

e-"'

501..,:r-~ /v'o r ·r A_

tt=cn<1fc>.

1500 . 00

f \ I

~/~

-1000 0~1---~~~-----~~--~~--~~----~-

-o 00~~~~~~~~---r~~~~-r-~~~~~~~-

0.00 s . 00 1-~~~~~-..-~--..~~c-::-:-.,....-~~~~~~~-<

5.00 D1 s tonc~ fKn1

10 00

Figure 2. Magnetic anomalies along (a) west-east and (b) south­north profiles, o~er a vertical prism Bhat nas a strong remanent magnetization (inc.rdec. = 40 ,290 ). *"'

The majority of the Bald Eagle ring troctolite samples used

for Table 1 were obtained along the eastern side of the

intrusion.

weakly magnetized rocks in the study area include

anorthositic, troctolitic, gabbroic and granophyric rocks.

The magnetic properties of these rocks are not generally

distinctive. If outcrop or drill hole information exists,

rock types may be assigned to weakly magnetized zones that

have been defined through modeling, filtering and imaging.

Rock magnetization measurements indicate that many of the

stronger magnetic anomalies observed over the Duluth Complex

are produced by troctolitic and gabbroic rocks that are

enriched in magnetite.

.fil;m.roach

The processing and interpretation approach applied ~n this

study utilizes three techniques, each having particular

strengths. These techniques include:

Werner Deconvolution-based inverse modeling.

- Talwani-based forward modeling.

- Frequency-domain filtering and data enhancement.

Werner Deconvolution

This inverse magnetic modeling technique is based on the

assumption that anomaly sources may be approximated by thin

sheets and interfaces of arbitrary dip and infinite strike

and depth extents (Fig. 3). The assumption of polynomial­

like anomaly interference allows a wide variety of geologic

features (Fig. 4) to be accurately characterized, even when

severe interference occurs.

Information provided by Werner deconvolution includes source

location (x 0 ), depth (z 0 ), dip (d) and susceptibility

contrast (k) estimates. These results are presented in map

5

Xo

t Zo

f Zo

THIN SHEET

-

INTERFACE

.. x

/ /

~ /

Figure 3. Ideal thin sheet and interface sources. The position, depth and dip of the stick symbols shown on the right-hand side correspond directly to those features of the actual sources. Thin sheets are indicated by circles, and interfaces by squares. Stick length is proportional to the reciprocal logarithm of the magnitude of k and is not related to the depth extent of the sheet. Negative k values are designated by a dashed stick and, for the case of an interface, indicate a susceptibility decrease in the positive x-<lirection.

6

THIN SHEET

W1-~r§i¥~t{~~

- I

"'I ___-€) - ~ -a

Figure 4 a . Geological features that can be approximated by thin sheets: (top) dike, (center) magnetic unit (i.e. iron-formation, etc.) contained in the limbs of a truncated fold, (bottom) magnetic lens (i.e. magnetic lava flow, pyrrhotite body, etc.) surrounded by non-magnetic rock. Corresponding stick symbols are shown on the right hand side.

INTERFACE

• - r w

-.:, ; . -·. __ . ';i)

~-----:._- .

~~-::: ~-- ---~--- -;

_.. I

~llJIIfa•t1~ ;:;.::.-'.,:::)fi.'::.~·,:i:;.:~;i::;..~·;::'.:. ~5Y.:t-f~j:· ~·~f~{t5:~:{{::,/,:~-~~~-+:~~?

(iJ I

' r

b

Figure 4 b . Geological features that can be approximated by dipping interfaces: (top) edges of a plutonic body, (center) leading edge of a thrust fault, (bottom) normal faults. Corresponding stick symbols are shown on the right hand side.

-...:

form so that they may be overlain on aerornagnetic maps. For

more detailed information and references, see Ferderer

(1988).

Because .geologic strike varies considerably in the study

area, the aerornagnetic data were gridded prior to

deconvolution. The resulting 100 m grid, was broken into

south-north and west-east oriented profiles which were

upward continued and deconvolved. When data are gridded,

smoothing occurs. As a result, gridding prior to

deconvolution may result in slight parameter estimate

errors. Upward continuation is used to minimize the effects

of noise.

Information was obtained for anomaly sources having a wide

variety of strike directions. The effects of geologic

strike on depth estimates were accounted for during across­

prof ile correlation of results.

Remanent magnetization was accounted for during

deconvolution.

Forward Modeling

Forward modeling was perfonned along four south-north, and

six west-east oriented grid lines. This technique assumes

uniformly magnetized, two-dimensional, polygonal-shaped

bodies, and is based on the technique of Talwani and

Heirtzler (1965), Remanent magnetization was accounted for

during modeling.

Forward modeling was also applied to one northwest-southeast

oriented Bouguer gravity profile.

Frequency Domain Filtering

Four fr~q~ency domain filtering techniques were applied to

the above mentioned 100 m grid. These techniques include:

- Reduction to the pole: The total magnetization vector,

comprised of induced and remanent magnetization

components, was reduced .

- Upward continuation: Data were calculated at levels of

350 m, 500 m and 1000 rn. Werner deconvolution was applied

to test profiles for each of these three levels, and the

500 m level was chosen for production processing.

- Second vertical derivative filter.

- Bandpass filters: Wavelengths less than or equal to 2, 5,

10 and 15 km were passed. The 15 km bandpass filter was

found to be quite useful for isolating shorter-wavelength

gravity anomalies. This technique was not heavily · relied

on for magnetic interpretation.

Filtering operations were performewl in the following

sequences:

1) Red. to pole

2) Red. to pole > upward contin. > 2nd vert. deriv.

3) Red. to pole > upward contin. > bandpass filter

4 ) Upward contin. > Werner deconvolution

Color and shaded relief maps were used to present the raw

and filtered data.

Interpretation

Results obtained using each of the three processing

techniques described above were plotted at a scale of

1:62,500.

The derived aeromagnetic interpretation is given on Plate 1.

9

Where Werner deconvolution solutions are strong, contacts

were drawn based on these results. In areas of weak or no

solutions, results of forward modeling and second vertical

derivative data w~re used to extrapolate contact positions.

Parameter estimates indicated on Plate 1 are average values,

obtained from groups of clustered Werner deconvolution

solutions. A small number of these estimates are based on

the results of forward modeling.

In most cases, interpreted faults are based on evidence

provided by Werner deconvolution. This technique is

especially useful for locating strike-slip faults, based on

sharp offsets in source position estimates. Normal faults

are often indicated by offsets in depth estimates .

Depth estimates represent the depth to a magnetic source.

This is not necessarily the depth to bedrock. In some

cases, deep estimates may represent moderately or strongly

magnetized intrusive rocks which lie below more weakly

magnetized rocks. Depth estimates are rounded to the

nearest 25 m for depths less than 100 m, and to the nearest

50 m for depths greater than 100 m.

In most cases, depth and dip estimates are assigned to

mapped contacts and layering. In a few instances, isolated

sets of parameter estimates which represent unmapped

structures, are given.

Discussion

The study area is located in the west-central portion of the

1.1 Ga Duluth Complex.

Bedrock throughout most of this area is poorly exposed, and

with the exception of the northwest corner and the northern

margin, mapping has only been performed at a regional scale

(1:250,000) (Green, 1982).

10

The South Kawishiwi Intrusion (SKI) occurs in the northwest

corner of the study area. The SKI is relatively well

exposed and has been studied by Foose and Cooper (1981), and

Foose and We i b l en (1986). On average, the dominantly

troctolitic rocks (ta) of the SKI strike N40-so 0 ~ and dip

between 5 and 20° SE.

Rocks of the SKI are weakly magnetized and have little

aeromagnetic expression. Measurements taken from these

rocks were included with those of the basal troctolite group

of Table 1.

A number of faults have been mapped in the SKI, only one of

which appears to be strongly reflected in the aeromagnetic

data. Four additional northwest trending faults that

intersect the SKI are mapped based on the results of this

study.

The Bald Eagle Intrusion (BEI) (Weiblen, 1965; Chandler,

1985) is a second, well studied structure occurring in ~he

study area. The southern end of this intrusion intersects

the northern margin of the study area. The BEI is a funnel­

shaped intrusion, consisting of a core gabbro surrounded by

troctolitic rocks.

Anomaly data suggest that the magnetizations and densities

of troctolitic rocks on the western and eastern sides of the

intrusion differ significantly. Troctolitic rocks on the

western side (tbl) are moderately magnetized and have a

relatively low density (two samples of this rock have a

density 2.73 gm/cc, Holst and others, 1986). Those on the

eastern side (tb2) are weakly magnetized and quite dense

(several samples have an average density of 3.03 gm/cc).

The magnetization of the core gabbro (gb) is intermediate to

those of the western and eastern troctolites, and it is also

quite dense.

11

A magnetic high occurs near the mapped contac t (Green and

others, 1966) between the core gabbro and the eastern

t roctolitic rocks. Model studies suggest that this high is

related to two factors, a zone of more magnetic gabbro or

troctolite (otgl), in contact with, or located near the

weakly magnetized eastern troctolitic rocks.

Immediately west of the BEI, rocks are weakly to moderately

magnetized. Both troctolitic and anorthositic rocks have

been mapped in this area by Green and others (1966).

Due to lack of exposure, the geology of the remainder of the

study area is much more poorly understood.

A unit denoted tsl, interpreted along the southeast margin

of the SKI and striking approximately N450E, is slightly

more magnetic than surrounding ts and t units. This unit

may be a reaction zone, produced when younger rocks were

intruded to the southeast of the SKI. Rocks of the t unit,

mapped just south of the SKI, may actually belong to the

SKI, in which case the tel unit would likely represent an

internal structure of the SKI.

Three large, oblong to circular shaped anomalies occur in

the western half of the study area. The units otgl-3 have

been assigned to the magnetic bodies that produce these

anomalies, referred to as the northwestern, southwestern and

eastern bodies. At least two of these bodies appear to be

magnetically zoned, which may reflect further petrologic

zoning. The three bodies are separated by rocks of the t

unit.

The weakly magnetized zone to which the t unit has been

assigned is widespread throughout the study area and more

poorly defined to the southeast. A small number of outcrops

indicate that rocks in this zone are troctolitic. Possible

relationships between the t unit and rocks that produce the

three large anomalies include the following:

1) A single, moderately to veJ~Y strongly magnetized

intrusion was forcibly injected by rocks of the t

unit, possibly along faults, and its parts were

spread apart.

2) Three separate intrusions of moderately to very

strongly magnetized rock were emplaced into the

weakly magnetized t unit.

3) The three anomalies are produced by large inclusions

of the Lower Proterozoic Biwabik Iron Formation, in

the t unit. This possibility seems least likely

based on geometries and magnetizations derived

through modeling.

The western and northern contacts of the northwestern body

are indicated to dip moderately to steeply to the northwest.

This is evidence that this body is discordant with rocks of

the SKI.

Modeling indicates that the southwestern body has a deep

root which may represent a feeder. It is further implied

that this root rises from the east and from depths in excess

of 5 km. Large depth estimates obtained along this anomaly

suggest that the mapped otgl unit is covered by weakly

magnetized rocks along the northern and western sides of the

anomaly.

Thin ledges of the otgl unit, which extend and thin to the

south, are interpreted for both the southwestern and eastern

bodies. These ledges ~ay have been shaped when the original

body or bodies were intruded by rocks of the t unit.

Alternatively, they may represent an original funnel shape

for the magnetic body or bodies.

Offsets in Werner deconvolution solutions indicate that

strike-slip faults occur within each of the three bodies.

13

Gravity and magnetic expressions just east of the eastern

body (near UTMs 606,5285) are especially interesting. The

anomaly source in this area is weakly magnetized, yet quite

dense. Forward modeling was applied to a residual Bouguer

gravity anomaly along a northwest-southeast trending profile

in this area. The residual anomaly was obtained by

subtracting values calculated assuming a linear gradient,

from data of Ikola (1970). The model suggests a source rock

density of 3.15 gm/cc, and it is implied that the source

rock contains weakly-magnetic oxides or considerable

olivine. Although these rocks appear to be denser than the

relatively dense troctolitic rocks of the eastern BEI to the

north, it is likely that the two rocks are related.

Accordingly, both of these dense rocks are assigned to the

unit tb2. Werner deconvolution results suggest that this

unit may form the core of the moderately magnetic eastern

body at depth.

Outcrops of anorthositic rocks (a) occur in the northwest,

northeast and southeast portions of the study area.

Consistent with rock property data, these rocks appear to be

weakly magnetized and have a low density. Their magnetic

expression is generally uninformative.

A unit denoted al, located in the northeast portion of the

study area is slightly more magnetic than surrounding rocks.

Anorthositic rocks outcrop on both sides of this unit near

the northern margin of the study area. It is therefore

likely that this unit reflects an original structure within

the anorthositic rocks.

East of the study area, a major eastward trending

aeromagnetic lineament splits into two lineaments which

transect the tau unit as they enter the northeast portion of

the study area, trending about N70oE. Strong Werner

deconvolution interface solutions, were obtained along the

northern lineament, and the anomaly source is interpreted to

be a fault. Magnetization decreases from south to north

across this fault, which appears to have been refaulted by

strike-slip faults that trend about N4s 0w.

Thin sheet solutions are s t rongest along the southern

lineament. This lineament is interpreted to correspond to a

second fault which has been intruded by moderately magnetic

material so that it generates a sheet-like anomaly. Both of

these lineaments are truncated by rocks in the west-central

portion of the study area, implying that they are older than

these rocks, or that they originated with the intrusion of

these rocks.

Several strong, elongate and plug shaped anomalies occur in

the western and southern parts of the study area. These

anomalies are interpreted to be produced by oxide-rich

troctolitic and gabbroic rocks, similar to those thought to

be responsible for the "Snake" or "Greenwood Lake ~omaly",

in the Greenwood Lake area (Vadis and others, 1981). It is

likely that the longest of these bodies is a continuation of

the source of the Snake Anomaly.

Dip estimates obtained along the contacts of these

intrusions indicate that they are wedge-shaped, their

contacts dipping outwards at moderate to steep angles.

Magnetizations estimated for these rocks are similar to

those for the rocks responsible for the three oblong to

circular shaped anomalies in western portion of the study

area. This does not necessarily imply a genetic

relationship.

Rocks occurring in the southeast portion of the study area

are magnetically indistinct. Anorthositic rocks are

observed at a small number of outcrops, and it is likely

that troctolitic rocks are also present in this area. The

unit tau is used to represent these rock~. It is also

15

likely that granophyric rocks are more widespread than is

indicated on Plate 1.

Forward modeling was performed along a south-north oriented

profile, near the southeastern corner of the study area.

The modeled bodies were all assigned depths greater than 500

m. This implies that more strongly magnetized rocks in the

southeast corner of the area are located at depth, beneath

less magnetic rocks.

Faulting does not always produce a magnetization contrast,

and it is likely that faulting is significantly more

widespread than indicated on Plate 1.

Re c orrune ndations

Drilling should be performed to test and refine the

interpretations that have been made here. Some of the

interpreted structures may have implications for mineral

exploration. Possible targets include:

The dense, weakly magnetized tb2 unit, along the

northeastern side of the large anomaly in the north­

central portion of the study area (near UTMs 606,5285).

It is likely that this unit does not intersect the bedrock

surface everywhere it has been mapped. Further

geophysical studies may be helpful in this regard.

- The series of intersecting faults interpreted near

(611,5284). Ground-geophysical surveys should be used to

more precisely locate these faults before drilling is

performed.

- The units tsl and al, to determine how they differ from

surrounding rocks.

- The unit t, at different locations, to more specifically

define its character.

lG

17

Rocks of the units otgl-3, which produce large, oblong to

circular shaped anomalies in the western half of the study

area, to verify that petrologic zoning occurs in these

rocks.

The analysis performed in this study is almost entirely

geophysical in nature. The resulting interpretation should

therefore be regarded as a first approximation of the

geology in the study area, and be used judiciously by

geologists working in the area.

Respectfully submitted,

;< 4~-t j . 9' ~~\. Robert J. Ferderer, Ph.D.

References Cited

Chandler, V.W., 1985, Interpretation of Precambrian geology in Minnesota using low-altitude, high-resolution aeromagnetic data, in Hinze, W.J., ed., The utility of regional gravity and magnetic anomaly maps: Tulsa, Oklahoma, Society of Exploration Geophysicists, p. 375-391.

Ferderer, R.J., 1982, Gravity and magnetic modeling of the southern half of the Duluth Complex, northeastern Minnesota: Unpublished M.S. thesis, Indiana University, Bloomington, 99 p.

Ferderer, R.J., 1988, Werner deconvolution and its application to the Penokean orogen, east-central Minnesota: Unpublished Ph.D. thesis, University of Minnesota, Minneapolis, 284 p.

Foose, M.P., and Cooper, R.W., 1981, Faulting and fracturing in part of the Duluth Complex, northeastern Minnesota: Canadian Journal of Earth Sciences, v. 18.

Foose, M.P., and Weiblen, P.W., 1986, The physical and petrologic setting and textural and compositional characteristics of sulfides from the South Kawishiwi Intrusion, Duluth Complex, Minnesota, USA: in Friedrich, G.H., ed., Geology and Metallogeny of Copper Deposits.

UJ

Green, J.C. , 19 8 2, Geologic Map of Minnesota, 'l'wo Harbors ~;;heet: Minnesota Geological Survey.

Green, J.C., Phinney, W.C., and Weiblen, P.W., 1966, Gabbro Lake quadrangle, Lake County, Minnesota: Minnesota Geological Survey Miscellaneous Map Series M-2.

Halls, H.C., and Pesonen, L.J., 1982, Paleomagnetism of Keweenawan rocks, in Wold, R.J., and Hinze, W.J., eds., Geology and Tectonics of the Lake Superior Basin: Geological Society of America Memoir 156, p. 173-202.

Holst, T.B., Mulleruneister, E.E., Chandler, V.W., Green, J.C., and Weiblen, P.W., 1986, Relationship of structural geology of the Duluth Complex to economic mineralization: Minnesota Dept. of Natural Resources Report 241-2.

Ikola, R.J., 1970, Simple Bouguer gravity map of Minnesota, Two Harbors sheet: Minnesota Geological Survey Miscellaneous Map Series M-9, scale 1:250,000.

Talwani, M., and Heirtzler, J.R., 1965, Computation of magnetic anomalies caused by two-dimensional structures of arbitrary shape: Computers in the Mineral Industries, part 1, Stanford University Publication, Geological Sciences, v. 9, p. 464-480.

Vadis, M.K., Gladen, L.W., and Meineke, D.G., 1981, Geological, geophysical, and geochemical surveys of Lake, St. Louis, and Cook Counties, Minnesota for the 1980 drilling project: Minnesota Dept. of Natural Resources, Division of Minerals Report 201, 13 p.

Weiblen, P.W., 1965, A funnel-shaped gabbro-troctolite intrusion in the Duluth Complex, Lake County, Minnesota: Unpublished Ph.D. dissertation, University of Minnesota, Minneapolis.